feat: Add nomic modern bert

mteb/models/nomic_models.py

@@ -7,6 +7,8 @@
 import numpy as np
 import torch
 import torch.nn.functional as F
+import transformers
+from packaging.version import Version
 
 import mteb
 from mteb.encoder_interface import PromptType
@@ -16,6 +18,8 @@
 
 logger = logging.getLogger(__name__)
 
+MODERN_BERT_TRANSFORMERS_MIN_VERSION = "4.48.0"
+
 
 class NomicWrapper(SentenceTransformerWrapper):
     """following the hf model card documentation."""
@@ -28,6 +32,14 @@ def __init__(
         **kwargs: Any,
     ):
         self.model_name = model_name
+        if model_name == "nomic-ai/modernbert-embed-base" and (
+            Version(transformers.__version__).release
+            < Version(MODERN_BERT_TRANSFORMERS_MIN_VERSION).release
+        ):
+            raise RuntimeError(
+                f"Current transformers version is {transformers.__version__} is lower than the required version"
+                f" {MODERN_BERT_TRANSFORMERS_MIN_VERSION}"
+            )
         super().__init__(model_name, revision, model_prompts, **kwargs)
 
     def to(self, device: torch.device) -> None:
@@ -92,7 +104,7 @@ def encode(  # type: ignore
 }
 
 nomic_embed_v1_5 = ModelMeta(
-    loader=partial(  # type: ignore
+    loader=partial(
         NomicWrapper,
         trust_remote_code=True,
         model_name="nomic-ai/nomic-embed-text-v1.5",
@@ -118,7 +130,7 @@ def encode(  # type: ignore
 )
 
 nomic_embed_v1 = ModelMeta(
-    loader=partial(  # type: ignore
+    loader=partial(
         NomicWrapper,
         trust_remote_code=True,
         model_name="nomic-ai/nomic-embed-text-v1",
@@ -144,7 +156,7 @@ def encode(  # type: ignore
 )
 
 nomic_embed_v1_ablated = ModelMeta(
-    loader=partial(  # type: ignore
+    loader=partial(
         NomicWrapper,
         trust_remote_code=True,
         model_name="nomic-ai/nomic-embed-text-v1-ablated",
@@ -171,7 +183,7 @@ def encode(  # type: ignore
 
 
 nomic_embed_v1_unsupervised = ModelMeta(
-    loader=partial(  # type: ignore
+    loader=partial(
         NomicWrapper,
         trust_remote_code=True,
         model_name="nomic-ai/nomic-embed-text-v1-unsupervised",
@@ -195,3 +207,31 @@ def encode(  # type: ignore
     adapted_from=None,
     superseded_by=None,
 )
+
+nomic_modern_bert_embed = ModelMeta(
+    loader=partial(
+        NomicWrapper,
+        model_name="nomic-ai/modernbert-embed-base",
+        revision="5960f1566fb7cb1adf1eb6e816639cf4646d9b12",
+        model_prompts=model_prompts,
+        model_kwargs={
+            "torch_dtype": torch.float16,
+        },
+    ),
+    name="nomic-ai/modernbert-embed-base",
+    languages=["eng-Latn"],
+    open_weights=True,
+    revision="5960f1566fb7cb1adf1eb6e816639cf4646d9b12",
+    release_date="2024-12-29",
+    n_parameters=149_000_000,
+    memory_usage=None,
+    max_tokens=8192,
+    embed_dim=768,
+    license="apache-2.0",
+    reference="https://huggingface.co/nomic-ai/modernbert-embed-base",
+    similarity_fn_name="cosine",
+    framework=["Sentence Transformers", "PyTorch"],
+    use_instructions=True,
+    adapted_from=None,
+    superseded_by=None,
+)